I have worked in newspapers and magazines for most of my career, so the use of these seems pretty ingrained to me. But it wasn’t always that way. I would suggest consulting a few resources, such as the links below. But many publications have their own particular styles for dealing with these issues, and you may have to adapt from job to job (such as the previous thread about “banned hyphenation.”)

http://desktoppub.about.com/cs/finetypography/ht/dashes_hyphens.htm

http://www.colorado.edu/Publications/styleguide/punctuation.html#dashes

[edit]
After rereading your question, I think I may not have uderstood what you were asking precisely. Sorry, if that’s the case.

If by sidebearings you mean a little breathing room on either side of the dash then yes, please!

The question of the appropriate side bearings of the em dash has been discussed here several times. Traditionally, at least in the US I believe it was as near to a full em as the metal type would allow, and was set with no additional spacing. Nowadays I believe it is usually designed with say 5% space on either side. Just checking one face, Minion has a em dash that is totals 922 units, including side bearings 50 either side—this is on a 1000 unit em. Thus the dash is about 90% of the advance width, with 5% side bearings on either side.

Kent Lew has explained that Matthew Carter favors an em dash that is 75% of the em, with the side bearings taking up the rest of the em. Kent has followed this practice, also negatively kerning the em dash with itself, so that a line of em dashes will form a continuous line. Note that Minion’s em dash is actually 822.

Carter’s practice makes a compromise between the British and American practices for parenthetical dashes. The British practice is to use en dashes with thin spaces either side. The American practice has been a full em with no extra spacing. Bringhurst rails against the American practice, decrying it as a relic of bad Victorian typography. Whatever the merit of this complaint, the current American practice moves toward the European style.

It shouldn’t be specific to a general and abstract editing principal, or a “good for all typefaces” great-man dictate, but depends on the typeface being used.

A high contrast serif face needs different dashing than a bold sans.

So this is certainly something that type designers can take care of, and typographers should be aware of how much “breathing room” the type designer/foundry has decided to include in the dash characters.

>A high contrast serif face needs different dashing than a bold sans.

The difference—aside from weight—being what?

>great-man dictate

I see from your new icon that you have acquired a halo. I guess we now need to see your comments as based on “a higher authority” :)

The difference—aside from weight—being what?

IMO, the lack of a gap between o and the dash is more acceptable in Bembo than Helvetica.

However, note how much space is built in to Electra — so the situation is wide open to interpretation, and typographers shouldn’t think that just because they follow an editing principal (spacing added, or not), that will produce consistent results with different faces.

you have acquired a halo

The Eye of God was already taken.

Nick —

The Eye of God was already taken.

ROFLMAO! That was priceless.

— K.

As Bill pointed out, for an em dash, the tradition in metal was a full em wide with the dash generally extending to the edges (i.e., zero sidebearings). Typesetters had different conventions (in conjunction with Editorial standards) of spacing these dashes. In the U.S., the most popular style is “closed” — that is, without space on either side. However, careful compositors would often insert a thin space to hold the dash away from rounds.

Nowadays approaches vary among type designers. I’ve outlined the thinking behind my approach in the notes of the Whitman PDF specimen at my site.

If you choose to outfit your em dash with positive sidebearings, then I recommend the kern that Bill mentioned above (at least for any serious text face). In bibliographies, it is common practice to use a 3-to-em dash to signify a repeated author entry, and this is usually set (not surprisingly) as 3 consecutive em dashes. If an emdash-emdash kern pair is not present, your unfortunate designer will find herself with a “dotted” 3-to-em dash (and will likely incur the wrath of the editors).

— K.

you have acquired a halo

Nick, should we address you as St. Nick from now on? ;-)

First make your own avatar, then mock those of others.

Yeah, yer right, mate. I used to have one, ya know... But then I took it down, planning to change it... and of course I haven’t gotten around to that yet!

Sincerely,
A Humbled Procrastinator ;-)

I’ve been using a 3/4 em dash on an em body (or dash 75% of the em width, with the rest of the space as sebearings) as long as Fontographer existed. I also make up a “twoemdash” and a “threeemdash” character with no sidebearings, so that I can get any length dash (2,3,4,5 ... n). In Type 1 fonts, you had to be able to write a custom encoding vector to use them, but if you could, it was a character assignment well spent.

With OpenType, you simply add them to your liga feature.

sub emdash emdash emdash by threeemdash
sub emdash emdash by twoemdash

When you think about it, the tradition of no sidebearings came about, at least in part, so you could set a two- or three-em dash. The large “case” of OpenType has changed that.

With the endash, you might want to give it the same sidebearings as the hyphen. While endashes are used more with numbers, both the hyphen and endash are used with both letters and numbers, and your kerning work is eased if they have the same sidebearings. In truth, you can make the sidebearings on the en about 5 to 10 units greater than the hyphen & still use the same kerning values; I think the overall fit with this is a bit better.

As for hyphen sidebearings, I make them to work with n m h etc. and use kerning for the fit with other letters. If your overall spacing of letters makes words with a certain characteristic (lose, tight), you will want that same characteristic with the hyphen.

FWIW

kentlew,

Glad to see the figuredash in Whitman. I have long felt that the n-dash, when used between uppercase numbers looks too low, and too tight on one side. I thought I was the only one . . .

The solution I have been using is to create a character style that raises the dash just enough.

Typical —

Well, that was a PS T1 hack. How quickly things change. The better solution now is probably to use the OT Layout {case} function to raise the en dash when all caps is applied to figures, yielding lining numerals. It might also be possible to use {calt} to implement an appropriate dash substitution for different fig styles.

— K.

@charles_e:
If you do have side bearings on em-dash than it’s far more easy and probably more right to kern the dash after a previous one to get rid of the space between.

All the rest:
THANK YOU! Just what I needed to know about en and em dashes.

There still is no info on hyphen width in terms of units or to em/en dash or any other letter. You said like: em dash = 3/4 of 1000... What about hyphen?

And the question if there’s a geometric sans serif typeface with slanted hyphen?

If you do have side bearings on em-dash than it’s far more easy and probably more right to kern the dash after a previous one to get rid of the space between.

Well, that’s why they race horses, I guess. But to me, they are different characters. Consider that a 3-em dash in a bibliography starts flush left. But you’re not going to kern an em-dash with a .125 em sidebearing flush left, at least, in InDesign. I’d rather use a different dash that starts without a left sidebearing, measures a full three ems, and, since it is usually followed by a period or comma, doesn’t have a .125 em right sidebearing. Bet on the gray if you want.

The two- and three-em dashes have various uses, such as to replace a persons name (e.g., Mrs.——), but all the uses I know of are different than the use of the em-dash.

Usually, the em-dash both separates and joins two somewhat related thoughts. Whether you think of it as “separating” or “joining” is at some level a word game, but graphically, the British word-spaced en dash follows the “separating” notion; the American closed-up em dash is in line with the “joining.” Some British designers who have worked in the States for a while, like Richard Eckersley, began to favor the 3/4-em dash in the full em space, and from this thread, it seems that Matthew Carter also signed on.

Another way of looking at an em-dash is that it replaces a pair of parentheses for expressing a thought that is related, but not grammatically connected to a sentence.

In InDesign (CS1) I will if necessary kern an em dash (or other character) that I want “flusher” left, into a preceding space. Has to be at the beginning of a new paragraph or after a forced line break, so works fine for bibliographies. I suppose a kern pair could be made in the font itself of space-em dash, though this would foul things up if putting a space-em dash somewhere else. Choice in a font of both em dash with no side bearings and em dash with them would be my desire.

Charles,

your “threeemdash” should have the glyph name “uni201420142014” or “emdash_emdash_emdash”, while your “twoendash” should have the glyph name “uni20132013” or “endash_endash”. Otherwise, they will not work properly in electronic documents.

All,

What should be the typographic treatment of ’FIGURE DASH’ (U+2012) and ’HORIZONTAL BAR’ (U+2015, formerly known as ’QUOTATION DASH’)?

A.